A method and system for evaluating the regulation potential of a regional data center cluster

Abstract

The application provides a regional data center cluster regulation potential evaluation method and system, and relates to the technical field of power system demand side management. The method comprises the following steps: collecting future load tasks of a regional data center cluster, and estimating power load demand by using a logarithmic model; quantitatively modeling the load correlation between data centers by using a Pearson correlation coefficient; comprehensively considering the load correlation and the energy consumption characteristics of each data center, and constructing a load transfer priority model to provide a direction for optimization; taking the minimum regional total power load demand as the target, integrating multiple constraints such as total load invariance, service quality, load upper limit and load transfer priority, constructing a regulation optimization model, and solving the model by using intelligent optimization algorithms such as particle swarm to obtain an optimal load allocation scheme; and calculating the regulation potential of the cluster by comparing the total power load demand before and after optimization. The accuracy and reliability of the evaluation of the regulation potential of the regional data center cluster are significantly improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of power system demand side management, in particular to a regional data center cluster regulation potential evaluation method and system. BACKGROUND

[0002] With the development of artificial intelligence technology, the computing demand of data centers in various enterprises and industries is increasing. Therefore, the number and capacity of data center loads in the power system are increasing, becoming an important part of the power system load. In addition, the computing load of data centers can be coordinated and distributed among regional data centers through communication networks, which is beneficial to improving the flexibility of power system operation. Therefore, in order to fully utilize the load regulation capability of regional data center clusters and support power grid operation optimization, an evaluation method capable of evaluating the regulation potential of regional data center clusters is urgently needed.

[0003] At present, the evaluation method of data center cluster regulation potential mainly comprehensively analyzes the energy consumption characteristics, load distribution and renewable energy utilization of data centers in multiple dimensions. Existing researches usually use mathematical modeling, machine learning algorithms and simulation techniques to evaluate the regulation capability of data centers under different operating conditions. However, these methods often ignore the interaction and correlation characteristics of load behaviors among multiple data centers in the region. Specifically, the load curves of different data centers may have a high positive correlation (peak at the same time) or a negative correlation (peak offset), and this correlation directly affects the effectiveness of load transfer strategies. If this correlation is ignored, the regulation potential evaluation result may be overly optimistic or pessimistic, which cannot provide accurate and reliable decision basis for power grid dispatching centers.

[0004] Therefore, there is an urgent need in the art for a method that can quantify the load correlation between regional data centers and accurately evaluate the overall regulation potential thereof. SUMMARY

[0005] The present application aims to overcome at least one technical problem in the prior art and provides a regional data center cluster regulation potential evaluation method and system.

[0006] In one aspect, the embodiment of the present application provides a regional data center cluster regulation potential evaluation method, which comprises the following steps: S1, collecting load task data of each data center in a region in a future set period and estimating power load of each data center; S2, calculating initial overall power load of all data centers in the region in the future set period based on the power load of each data center; S3, calculating a Pearson correlation coefficient of load curves of any two data centers in the region in the set period to quantify load correlation between the any two data centers; S4, constructing a data center load transfer priority model based on the load correlation between the any two data centers and energy consumption characteristic coefficients of the two data centers; S5, calculating a load transfer amount from any source data center to any target data center based on the data center load transfer priority model; S6, taking minimization of overall power load demand of the regional data center cluster as an objective function and taking the load transfer amount from any source data center to any target data center calculated based on the data center load transfer priority model as a guide, constructing a regional data center cluster regulation optimization model containing at least one actual constraint; S7, solving the regional data center cluster regulation optimization model by using a preset intelligent optimization algorithm to obtain an optimized load distribution scheme of each data center; S8, calculating an optimized overall power load of all data centers in the region in the future set period based on the optimized load distribution scheme of each data center; S9, calculating a regulation potential of the regional data center cluster based on the optimized overall power load of all data centers in the region in the future set period and the initial overall power load of all data centers in the region in the future set period.

[0007] Further, the step S1 comprises: estimating the power load of each data center by using a logarithmic model based on a data center energy consumption characteristic coefficient and real-time load of the data center; and a mathematical expression of the logarithmic model is:

[0008] ;

[0009] A mathematical expression of the initial overall power load of all data centers in the region in the future set period is:

[0010] ;

[0011] In the formula, Pi is the estimated power load of the i th data center, and the unit is MW; Pi, nom is the known rated power of the i th data center, and the unit is MW; Ci is the energy consumption characteristic coefficient of the i th data center, and the value range is 0.8 to 2.0; This represents the real-time load of the i-th data center, expressed in times per second. is the load limit for the i-th data center, in times / s; n is the number of data centers in the region; Set the initial overall power load for all data centers in the region for the future period, in MW.

[0012] Furthermore, the formula for calculating the Pearson correlation coefficient in step S3 is as follows:

[0013] ;

[0014] In the formula, This represents the Pearson correlation coefficient between the load curves of data centers i and j over a future set period T. Let this represent the load of data center i at time t within a future set period T. This represents the load of data center j at time t within a future set period T; This represents the average load value of data center i over a future set period T. Let t represent the average load value of data center j over a future set period T; t represents the time within the future set period T.

[0015] Furthermore, the mathematical expression for the data center load balancing priority model in step S4 is:

[0016] ;

[0017] ;

[0018] In the formula, This represents the load transfer amount for data center i. This represents the load transferred from data center i to data center j; This is the load transfer factor; Let be the energy consumption characteristic coefficient of data center j. It is represented as the energy consumption characteristic coefficient of data center s.

[0019] Furthermore, the objective function in step S6 is:

[0020] ;

[0021] In the formula, This represents the optimized overall power load demand of the regional data center cluster, expressed in MW.

[0022] Furthermore, at least one practical constraint in step S6 includes: a constraint that the load of all data centers in the region remains unchanged before and after optimization, and the total load remains unchanged; a service quality constraint that the load transmission delay between any two data centers shall not exceed the set delay limit; a load limit constraint that the load of each data center after optimization shall not exceed its load limit; and a load transfer priority constraint that the load transfer direction and proportion are guided by the data center load transfer priority model.

[0023] Furthermore, the total load constant constraint is: ;

[0024] The service quality constraints are as follows: ;

[0025] The load limit constraint is: ;

[0026] In the formula, n represents the number of regional data centers. This represents the load transfer latency between data centers z and i. The upper limit of transmission latency required by quality of service constraints. and These represent the real-time load and load limit of data center i, respectively. This represents the load transfer amount for data center z. This represents the amount of load transfer from data center z to data center i.

[0027] Furthermore, the preset intelligent optimization algorithm in step S7 is the particle swarm optimization algorithm, and the target variable for optimization calculation is the load transfer amount of each data center.

[0028] Furthermore, step S9 includes:

[0029] ;

[0030] In the formula, This represents the controllability potential of a regional data center cluster, expressed in MW. This represents the initial total power load of all data centers within the region over a future specified period, in MW. This represents the optimized overall power load demand of the regional data center cluster, expressed in MW.

[0031] Secondly, embodiments of the present invention provide a regional data center cluster control potential assessment system. The system employs the aforementioned regional data center cluster control potential assessment method. The system includes: a data center power load estimation module, adapted to collect load task data of each data center within a future set period and estimate the power load of each data center; a regional initial power load estimation module, adapted to calculate the initial overall power load of all data centers within the region within a future set period based on the power load of each data center; a load correlation calculation module, adapted to calculate the Pearson correlation coefficient of the load curves of any two data centers within the set period to quantify the load correlation between any two data centers within the region; a data center load transfer priority model construction module, adapted to construct a data center load transfer priority model based on the load correlation between any two data centers and the energy consumption characteristic coefficients of the two data centers; and a load transfer amount calculation module, adapted to calculate the load transfer amount based on the data center load transfer priority model. The system includes: a model for calculating the load transfer from any source data center to any target data center; a regional data center cluster regulation and optimization model construction module, suitable for constructing a regional data center cluster regulation and optimization model with at least one practical constraint, guided by the data center load transfer priority model, with the goal of minimizing the overall power load demand of the regional data center cluster; a load allocation scheme generation module, suitable for solving the regional data center cluster regulation and optimization model using a preset intelligent optimization algorithm to obtain the optimized load allocation scheme for each data center; an optimized regional power load estimation module, suitable for calculating the overall power load of all data centers in the region within a future set period based on the optimized load allocation scheme for each data center; and a regional data center cluster regulation potential calculation module, suitable for calculating the regulation potential of the regional data center cluster based on the overall power load of all data centers in the region within a future set period and the initial overall power load of all data centers in the region within a future set period.

[0032] Thirdly, embodiments of the present invention also provide an electronic device, including: a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the computer program, when executed by the processor, implements the above-described method for assessing the control potential of a regional data center cluster.

[0033] Fourthly, embodiments of the present invention also provide a readable storage medium, which, when the instructions in the storage medium are executed by the processor of an electronic device, enables the electronic device to perform the above-described regional data center cluster control potential assessment method.

[0034] The advantages of this invention compared to the prior art are as follows: Compared to the prior art, the advantages of this invention are:

[0035] 1. High assessment accuracy: This invention incorporates the load correlation between regional data centers into the control potential assessment framework, and accurately quantifies the spatiotemporal complementary characteristics of the load curves through the Pearson coefficient, avoiding assessment bias caused by ignoring correlation.

[0036] 2. Strong strategy orientation: The constructed load transfer priority model combines abstract correlation coefficients with specific energy consumption characteristics, providing a clear and efficient search direction for the optimization algorithm, so that the load transfer strategy can not only reduce the total energy consumption, but also conform to the business operation rules.

[0037] 3. High practicality: The optimization model fully considers actual engineering constraints such as quality of service, transmission delay, and load capacity, ensuring the feasibility and practicality of the evaluation results and the resulting scheduling schemes, and can directly provide reliable decision support for power grid dispatching departments.

[0038] 4. High degree of automation: This method forms a complete automated process from data input to potential assessment output, requiring minimal human intervention, resulting in high assessment efficiency and suitability for online analysis and real-time scheduling scenarios. Attached Figure Description

[0039] The present invention will be further described below with reference to the accompanying drawings and embodiments.

[0040] Figure 1 This is a flowchart of a method for assessing the control potential of a regional data center cluster, provided in Embodiment 1 of the present invention.

[0041] Figure 2 This is a schematic diagram of the IEEE 14-node test system structure used in Embodiment 1 of the present invention.

[0042] Figure 3 This is a schematic diagram of the daily load curves of the three data centers provided in Embodiment 1 of the present invention.

[0043] Figure 4 This is a schematic diagram of the daily power load demand of each data center based on the original load estimation provided in Embodiment 1 of the present invention.

[0044] Figure 5 This is a schematic diagram of the convergence curve of the utility function when solving an optimization model using the particle swarm optimization algorithm, as provided in Embodiment 1 of the present invention.

[0045] Figure 6 This is a schematic diagram of the assessment results of the regional data center cluster regulation potential provided in Embodiment 1 of the present invention.

[0046] Figure 7 This is a schematic diagram of a regional data center cluster regulation potential assessment system provided in Embodiment 2 of the present invention.

[0047] Figure 8 This is a partial block diagram of the electronic device provided in Embodiment 3 of the present invention. Detailed Implementation

[0048] Before discussing the exemplary embodiments in more detail, it should be noted that some exemplary embodiments are described as processes or methods depicted as flowcharts. Although the flowcharts describe the operations as sequential processes, many of these operations can be performed in parallel, concurrently, or simultaneously. Furthermore, the order of the operations can be rearranged. The process can be terminated when its operation is completed, but may also have additional steps not included in the figures. The process can correspond to a method, function, procedure, subroutine, subroutine, etc.

[0049] It should be understood that although the terms "first," "second," etc., may be used herein to describe various units, these units should not be limited by these terms. These terms are used merely to distinguish one unit from another. For example, without departing from the scope of the exemplary embodiments, a first unit may be referred to as a second unit, and similarly, a second unit may be referred to as a first unit. The term "and / or" as used herein includes any and all combinations of one or more of the associated listed items.

[0050] The present invention will now be described in detail with reference to the accompanying drawings. These drawings are simplified schematic diagrams, illustrating only the basic structure of the invention, and therefore only show the components relevant to the invention.

[0051] Example 1

[0052] The specific implementation method is as follows:

[0053] like Figure 1 The diagram shown is a flowchart of a method for assessing the control potential of a regional data center cluster provided by the present invention.

[0054] As an example, the method includes: Step S1, collecting load task data of each data center in the region within a future set period, and estimating the power load of each data center; Step S2, calculating the initial overall power load of all data centers in the region within the future set period based on the power load of each data center; Step S3, calculating the Pearson correlation coefficient of the load curves of any two data centers in the region within the set period to quantify the load correlation between any two data centers in the region; Step S4, constructing a data center load transfer priority model based on the load correlation between any two data centers and the energy consumption characteristic coefficients of the two data centers; Step S5, calculating the load transfer amount from any source data center to any target data center based on the data center load transfer priority model; Step S6, using the regional data center cluster... The objective function is to minimize the overall power load demand. Guided by the load transfer amount from any source data center to any target data center calculated by the data center load transfer priority model, a regional data center cluster regulation and optimization model containing at least one practical constraint is constructed. Step S7: The regional data center cluster regulation and optimization model is solved using a preset intelligent optimization algorithm to obtain the optimized load allocation scheme for each data center. Step S8: Based on the optimized load allocation scheme for each data center, the overall power load of all data centers in the region within a future set period is calculated. Step S9: Based on the overall power load of all data centers in the region within a future set period and the initial overall power load of all data centers in the region within a future set period, the regulation potential of the regional data center cluster is calculated.

[0055] In some feasible implementations, step S1 includes: estimating the power load of each data center using a logarithmic model based on the data center energy consumption characteristic coefficient and the real-time load of the data center;

[0056] The mathematical expression for the logarithmic model is:

[0057] ;

[0058] The mathematical expression for the initial overall power load of all data centers in the region within a future set period is:

[0059] ;

[0060] In the formula, The estimated power load of the i-th data center is given in MW. The known rated power of the i-th data center is expressed in MW. Let be the energy consumption characteristic coefficient of the i-th data center, with a value ranging from 0.8 to 2.0; This represents the real-time load of the i-th data center, expressed in times per second. is the load limit for the i-th data center, in times / s; n is the number of data centers in the region; Set the initial overall power load for all data centers in the region for the future period, in MW.

[0061] In some feasible implementations, the formula for calculating the Pearson correlation coefficient in step S3 is:

[0062] ;

[0063] In the formula, This represents the Pearson correlation coefficient between the load curves of data centers i and j over a future set period T. Let this represent the load of data center i at time t within a future set period T. This represents the load of data center j at time t within a future set period T; This represents the average load value of data center i over a future set period T. Let t represent the average load value of data center j over a future set period T; t represents the time within the future set period T. The future set period T is typically set to 24 hours.

[0064] Preferred, The value range of is [-1, 1]. When When data center i and data center j are perfectly positively correlated, in short, when data center i is busy, data center j is also busy; when When data center i and data center j are completely negatively correlated, in short, when data center i is busy, data center j is idle; when When the correlation coefficient is negative, it means that data centers i and j are uncorrelated; in short, their workloads are unrelated. By calculating the Pearson correlation coefficient, potential opportunity windows for task migration can be identified. The stronger the negative correlation, the greater the potential benefit of migrating the workload.

[0065] In some feasible implementations, the mathematical expression for the data center load balancing priority model in step S4 is:

[0066] ;

[0067] ;

[0068] In the formula, This represents the load transfer amount for data center i. This represents the load transferred from data center i to data center j; This is the load transfer factor; Let be the energy consumption characteristic coefficient of data center j. This is represented by the energy consumption characteristic coefficient of data center s. That is, when the amount of load that data center i needs to transfer is known, the load that data center i needs to transfer can be transferred to another data center with the strongest negative correlation with it by using the data center load transfer priority model and combining it with the Pearson correlation coefficient.

[0069] In some feasible implementations, the objective function in step S6 is:

[0070] ;

[0071] In the formula, This represents the optimized overall power load demand of the regional data center cluster, expressed in MW.

[0072] In some feasible implementations, at least one practical constraint in step S6 includes: a constraint that the load of all data centers in the region remains unchanged before and after optimization and the total load remains unchanged; a service quality constraint that the load transmission delay between any two data centers shall not exceed a set delay limit; a load limit constraint that the load of each data center after optimization shall not exceed its load limit; and a load transfer priority constraint that the load transfer direction and proportion are guided by a data center load transfer priority model.

[0073] Preferably, the total load constant constraint is: ;

[0074] The service quality constraints are as follows: ;

[0075] The load limit constraint is: ;

[0076] In the formula, n represents the number of regional data centers. This represents the load transfer latency between data centers z and i. The upper limit of transmission latency required by quality of service constraints. and These represent the real-time load and load limit of data center i, respectively. This represents the load transfer amount for data center z. This represents the amount of load transfer from data center z to data center i.

[0077] In some feasible implementations, the preset intelligent optimization algorithm in step S7 is the particle swarm optimization algorithm, and the target variable for optimization calculation is the load transfer amount of each data center.

[0078] Preferably, since the particle swarm optimization algorithm is already very mature in existing technology, its optimal solution process will not be described in detail here. The solution process can be summarized as follows: Step 1: Encoding. Represent a possible solution (i.e., a load allocation scheme) as the position of a "particle" in the particle swarm; Step 2: Initialization. Randomly generate a group of particles, whose initial positions can be guided by the load transfer priority model; Step 3: Iterative optimization. a. Evaluation: Calculate the fitness value of each particle (i.e., the total power load under this scheme). Particles that violate constraints will be "penalized" (their fitness deteriorates); b. Update: Each particle updates its velocity and position based on its own historical best position and the group's historical best position, moving closer to a better solution; c. Convergence: Repeat the iteration until the fitness function no longer improves significantly or the maximum number of iterations (e.g., 500 times) is reached; Step 4: Output. Output the optimal position found by the swarm as the optimal load allocation scheme.

[0079] In some feasible implementations, step S9 includes:

[0080] ;

[0081] In the formula, This represents the controllability potential of a regional data center cluster, expressed in MW. This represents the initial total power load of all data centers within the region over a future specified period, in MW. This represents the optimized overall power load demand of the regional data center cluster, expressed in MW.

[0082] To facilitate understanding of the above implementation methods, specific examples are provided here:

[0083] This embodiment is based on the IEEE 14-node system, as shown in the attached diagram. Figure 2 As shown in the figure, assuming there are 3 data centers at nodes 4, 6, and 8, their loads at different times are as follows. Figure 3 As shown, the maximum load rates are 240,000 times / s, 40,000 times / s and 240,000 times / s respectively, and the data center energy consumption coefficients are 1.0, 1.0 and 1.0 respectively, the rated power is 1MW, and the information transmission delay between data centers is 20ms.

[0084] First, based on the load of the three data centers, calculate the power load demand of each data center, and then calculate the initial total power load (baseline value before optimization):

[0085] According to the power load estimation formula in step S1: The load of each individual data center is calculated separately, and the power load values ​​of each data center at different times during the next day are as follows: Figure 4As shown.

[0086] Secondly, the correlation between the three data centers was analyzed using the Pearson correlation coefficient, and the results are shown in Table 1 below:

[0087] Table 1: Data Center Load Correlation Analysis

[0088]

[0089] It is evident that there is a complete negative correlation between Data Center 1 and Data Center 3.

[0090] Then, based on the energy consumption coefficient and load correlation of the data centers, a load transfer priority model was designed, and a regional data center cluster regulation and optimization model was constructed. The particle swarm optimization algorithm was used for calculation, with the load adjustment amount of each data center as the state variable. The fitness function changes are shown in the attached figure. Figure 5 As shown in the figure, the optimal data center load allocation scheme is obtained after 500 iterations of calculation.

[0091] Finally, by combining the optimized overall power load of all data centers in the region over a future set period with the initial overall power load of all data centers in the region over a future set period, the controllability potential of the regional data center cluster is calculated, as shown in the appendix. Figure 6 As shown in the figure, the potential for regulating the total power demand of the regional data center cluster during the day can be seen.

[0092] The above implementation addresses the problems of existing technologies: data centers are typically treated as independent entities for energy consumption and potential estimation, or simply considered as superposition. This approach completely ignores the interaction of load behaviors between data centers (such as "simultaneous busy / idle" or "peak shaving and valley filling"), leading to assessment results that deviate significantly from reality. It either overestimates potential (ignoring the inability to reduce synchronous peaks caused by positive correlation) or underestimates potential (failing to discover the significant optimization space brought about by negative correlation). This implementation creatively introduces "load correlation" as a core evaluation dimension, quantifying the dynamic relationship between data centers through the Pearson correlation coefficient. It can accurately identify which data center combinations are "golden partners" (strong negative correlation), providing effective optimization paths for load transfer. Simultaneously, it can identify which combinations are futile for load transfer (strong positive correlation), avoiding the optimization algorithm searching in ineffective directions and improving efficiency and accuracy. The final regulatory potential assessment result is based on a deep understanding of its internal dynamic relationships, thus achieving a qualitative leap in reliability and accuracy, providing a more credible decision-making basis for power grid dispatch.

[0093] Compared to existing technologies, which suffer from problems such as slow convergence and susceptibility to local optima due to the lack of clear guidance when constructing optimization models, this invention constructs a load transfer priority model based on energy consumption characteristics and correlations. This model provides a clear "navigation map" for the optimization algorithm. During population initialization and iterative updates, the algorithm prioritizes paths with high load transfer coefficients, significantly accelerating convergence to the global optimum (e.g., ...). Figure 5 (The convergence curve shown). This ensures that the final load allocation scheme is not only mathematically optimal, but also a high-quality strategy that conforms to business logic and energy efficiency principles.

[0094] Compared to existing technologies, many research optimization models are overly idealistic, ignoring the hard constraints of real-world engineering, resulting in optimal solutions that cannot be implemented in practice. This implementation addresses this by systematically integrating multiple key practical constraints into the regulation optimization model, including constant total load, quality of service (transmission latency), and load limits. This ensures business continuity: the constant total load constraint guarantees the complete execution of all user computing tasks. It also guarantees user experience: the quality of service constraint ensures that load shifting does not introduce unacceptable network latency, preventing the sacrifice of service quality for energy saving. Furthermore, it ensures system security: the load limit constraint prevents any data center from experiencing performance bottlenecks or downtime due to overload. Finally, the conclusions are practical: the final evaluated regulation potential is an achievable potential under strict satisfaction of all real-world constraints, rather than a theoretical maximum value. The evaluation conclusions have direct and significant guiding value for power grids and enterprises.

[0095] Compared to existing technologies, which often rely on expert experience or manual configuration, making them ill-suited to dynamically changing loads and grid demands, this implementation automates the entire process from data input and correlation analysis to model building and optimization. It can rapidly respond to grid dispatch commands, performing near real-time or day-ahead potential assessments, thus meeting the demands of modern power systems for demand-side response speed. Furthermore, the framework of this method is highly scalable, allowing for easy integration of more data centers into the cluster for assessment and the incorporation of new constraints (such as renewable energy output and electricity price signals), broadening its application scenarios.

[0096] In summary, this invention incorporates load correlation, a key but often overlooked factor, into the analytical framework and constructs a complete methodology that balances theoretical depth with engineering practice. Ultimately, it achieves a more accurate, reliable, and practical assessment of the control potential of regional data center clusters, providing a solid technical foundation for fully exploring and utilizing this important flexible power grid resource.

[0097] Example 2

[0098] Please see Figure 7 This embodiment provides a schematic diagram of a regional data center cluster control potential assessment system.

[0099] As an example, the system employs the regional data center cluster control potential assessment method described in Example 1, and the system includes:

[0100] The Data Center Power Load Estimation Module 700 is suitable for collecting load task data of each data center in a region within a future set period and estimating the power load of each data center.

[0101] The regional initial power load estimation module 710 is suitable for calculating the initial overall power load of all data centers in the region within a future set period based on the power load of each data center.

[0102] The load correlation calculation module 720 is suitable for calculating the Pearson correlation coefficient of the load curves of any two data centers within a set period, so as to quantify the load correlation between any two data centers within the region.

[0103] The data center load transfer priority model construction module 730 is suitable for constructing a data center load transfer priority model based on the load correlation between any two data centers and the energy consumption characteristic coefficients of the two data centers.

[0104] The load transfer calculation module 740 is suitable for calculating the load transfer amount from any source data center to any target data center based on the data center load transfer priority model.

[0105] The regional data center cluster regulation and optimization model construction module 750 is suitable for constructing a regional data center cluster regulation and optimization model that includes at least one practical constraint, with the objective function of minimizing the overall power load demand of the regional data center cluster and guided by the data center load transfer priority model.

[0106] The load allocation scheme generation module 760 is suitable for solving the regional data center cluster control optimization model using a preset intelligent optimization algorithm to obtain the optimized load allocation scheme for each data center.

[0107] The optimized regional power load estimation module 770 is suitable for calculating the overall power load of all data centers in the region within a future set period based on the optimized load allocation scheme of each data center.

[0108] The regional data center cluster regulation potential calculation module 780 is suitable for calculating the regulation potential of the regional data center cluster based on the overall power load of all data centers in the region within a future set period after optimization and the initial overall power load of all data centers in the region within a future set period.

[0109] It is not difficult to see that this embodiment is a system implementation corresponding to the first embodiment, and this embodiment can be implemented in conjunction with the first embodiment. The relevant technical details mentioned in the first embodiment are still valid in this embodiment, and will not be repeated here to reduce repetition. Accordingly, the relevant technical details mentioned in this embodiment can also be applied to the first embodiment.

[0110] It is worth mentioning that all modules involved in this embodiment are logical units. In practical applications, a logical unit can be a physical unit, a part of a physical unit, or a combination of multiple physical units. Furthermore, to highlight the innovative aspects of this invention, this embodiment does not introduce units that are not closely related to solving the technical problem proposed by this invention; however, this does not mean that other units are absent from this embodiment.

[0111] Example 3

[0112] Please see Figure 8 The present invention also provides an electronic device, including: a memory and a processor; the memory stores at least one program instruction; the processor loads and executes the at least one program instruction to implement the regional data center cluster control potential assessment method provided in Embodiment 1.

[0113] The memory 802 and processor 801 are connected via a bus, which may include any number of interconnecting buses and bridges, connecting various circuits of one or more processors 801 and memory 802 together. The bus may also connect various other circuits, such as peripheral devices, voltage regulators, and power management circuits, which are well known in the art and therefore will not be described further herein. A bus interface provides an interface between the bus and the transceiver. The transceiver may be a single element or multiple elements, such as multiple receivers and transmitters, providing a unit for communicating with various other devices over a transmission medium. Data processed by processor 801 is transmitted over a wireless medium via an antenna, which further receives data and transmits it to processor 801.

[0114] The processor 801 is responsible for managing the bus and general processing, and can also provide various functions, including timing, peripheral interfaces, voltage regulation, power management, and other control functions. The memory 802 can be used to store data used by the processor 801 during operation.

[0115] Example 4

[0116] This invention also proposes a storage medium storing a method for assessing the control potential of a regional data center cluster. When the regional data center cluster control potential assessment program is executed by a processor, it implements the steps of the method described above. Since this storage medium employs all the technical solutions of the above embodiments, it possesses at least all the beneficial effects brought about by the technical solutions of the above embodiments, which will not be elaborated upon further here.

[0117] The above descriptions are merely embodiments of the present invention. Commonly known structures and characteristics are not described in detail here. Those skilled in the art are aware of all common technical knowledge in the field prior to the application date or priority date, are aware of all existing technologies in that field, and have the ability to apply conventional experimental methods prior to that date. Those skilled in the art can, based on the guidance provided in this application, improve and implement this solution in combination with their own capabilities. Some typical known structures or methods should not be obstacles for those skilled in the art to implement this application. It should be noted that those skilled in the art can make several modifications and improvements without departing from the structure of the present invention. These should also be considered within the scope of protection of the present invention, and will not affect the effectiveness of the implementation of the present invention or the practicality of the patent. The scope of protection claimed in this application should be determined by the content of its claims, and the specific embodiments described in the specification can be used to interpret the content of the claims.

Claims

1. A method for assessing the control potential of a regional data center cluster, characterized in that, The method includes: Step S1: Collect load task data of each data center in the region within a future set period, and estimate the power load of each data center; Step S2: Calculate the initial overall power load of all data centers in the region within a future set period based on the power load of each data center; Step S3: Calculate the Pearson correlation coefficient of the load curves of any two data centers within the set period to quantify the load correlation between any two data centers within the region. The formula for calculating the Pearson correlation coefficient is as follows: ； In the formula, This represents the Pearson correlation coefficient between the load curves of data centers i and j over a future set period T. Let this represent the load of data center i at time t within a future set period T. This represents the load of data center j at time t within a future set period T; This represents the average load value of data center i over a future set period T. Let be the average load value of data center j over a future set period T; t represents the time within the future set period T. Step S4: Construct a data center load transfer priority model based on the load correlation between any two data centers and the energy consumption characteristic coefficients of the two data centers. The mathematical expression of the data center load transfer priority model is: ； ； In the formula, This represents the load transfer amount for data center i. This represents the load transferred from data center i to data center j; This is the load transfer factor; Let be the energy consumption characteristic coefficient of data center j. This is represented by the energy consumption characteristic coefficient of data center s; Step S5: Calculate the load transfer amount from any source data center to any target data center based on the data center load transfer priority model; Step S6: Taking the minimization of the overall power load demand of the regional data center cluster as the objective function, and guided by the load transfer amount from any source data center to any target data center calculated by the data center load transfer priority model, construct a regional data center cluster control optimization model containing at least one practical constraint. The at least one practical constraint includes: the load of all data centers in the region before and after optimization and the total load remaining unchanged; the service quality constraint that the load transmission delay between any two data centers shall not exceed the set delay limit; the load limit constraint that the load of each data center after optimization shall not exceed its load limit; and the load transfer priority constraint that the load transfer direction and proportion are guided by the data center load transfer priority model. The constraint that the total load remains constant is: ; The service quality constraints are as follows: ; The load limit constraint is: ; In the formula, n represents the number of regional data centers. This represents the load transfer latency between data centers z and i. The upper limit of transmission latency required by quality of service constraints. and These represent the real-time load and load limit of data center i, respectively. This represents the load transfer amount for data center z. This represents the amount of load transfer from data center z to data center i; Step S7: Use a preset intelligent optimization algorithm to solve the regional data center cluster control optimization model to obtain the optimized load allocation scheme for each data center. Step S8: Calculate the overall power load of all data centers in the region within a future set period based on the optimized load distribution scheme of each data center. Step S9: Calculate the control potential of the regional data center cluster based on the overall power load of all data centers in the region within a future set period after optimization and the initial overall power load of all data centers in the region within a future set period.

2. The method for assessing the control potential of a regional data center cluster according to claim 1, characterized in that, Step S1 includes: The power load of each data center is estimated using a logarithmic model based on the data center energy consumption characteristic coefficient and the real-time load of the data center. The mathematical expression for the logarithmic model is: ； The mathematical expression for the initial overall power load of all data centers in the region within a future set period is: ； In the formula, The estimated power load of the i-th data center is given in MW. The known rated power of the i-th data center is expressed in MW. Let be the energy consumption characteristic coefficient of the i-th data center, with a value ranging from 0.8 to 2.0; This represents the real-time load of the i-th data center, expressed in times per second. is the load limit for the i-th data center, in times / s; n is the number of data centers in the region; Set the initial overall power load for all data centers in the region for the future period, in MW.

3. The method for assessing the control potential of a regional data center cluster according to claim 2, characterized in that, The objective function in step S6 is: ； In the formula, This represents the optimized overall power load demand of the regional data center cluster, expressed in MW.

4. The method for assessing the control potential of a regional data center cluster according to claim 1, characterized in that, The preset intelligent optimization algorithm in step S7 is the particle swarm optimization algorithm, and the target variable for optimization calculation is the load transfer amount of each data center.

5. The method for assessing the control potential of a regional data center cluster according to claim 1, characterized in that, Step S9 includes: ； In the formula, This represents the controllability potential of a regional data center cluster, expressed in MW. This represents the initial total power load of all data centers within the region over a future specified period, in MW. This represents the optimized overall power load demand of the regional data center cluster, expressed in MW.

6. A system for assessing the control potential of a regional data center cluster, wherein the system employs the method for assessing the control potential of a regional data center cluster as described in any one of claims 1-5, characterized in that, The system includes: The data center power load estimation module is suitable for collecting load task data of each data center in a region within a future set period and estimating the power load of each data center. The regional initial power load estimation module is suitable for calculating the initial overall power load of all data centers in the region within a future set period based on the power load of each data center. The load correlation calculation module is suitable for calculating the Pearson correlation coefficient of the load curves of any two data centers within a set period, so as to quantify the load correlation between any two data centers within the region. The data center load transfer priority model construction module is suitable for constructing a data center load transfer priority model based on the load correlation between any two data centers and the energy consumption characteristic coefficients of the two data centers. The load transfer calculation module is suitable for calculating the load transfer amount from any source data center to any target data center based on the data center load transfer priority model. The regional data center cluster regulation and optimization model construction module is suitable for constructing a regional data center cluster regulation and optimization model with at least one practical constraint, taking the minimization of the overall power load demand of the regional data center cluster as the objective function and the load transfer amount from any source data center to any target data center calculated by the data center load transfer priority model as the guide. The load allocation scheme generation module is suitable for solving the regional data center cluster control optimization model using a preset intelligent optimization algorithm to obtain the optimized load allocation scheme for each data center. The optimized regional power load estimation module is suitable for calculating the overall power load of all data centers in the region within a future set period based on the optimized load allocation scheme of each data center. The regional data center cluster regulation potential calculation module is applicable to calculating the regulation potential of the regional data center cluster based on the overall power load of all data centers in the region within a future set period after optimization and the initial overall power load of all data centers in the region within a future set period.

Images